APPLIED PHYSICS LETTERS 99, 152105 (2011)
Electrical-field induced giant magnetoresistivity in (non-magnetic)
phase change films
Junji Tominaga,1,a) Robert E. Simpson,1 Paul Fons,1,2 and Alexander V. Kolobov1,2
1
Green Nanoelectronics Center (GNC), Nanoelectronics Research Institute, National Institute of Advanced
Industrial Science & Technology (AIST), Tsukuba Central 4, Higashi 1-1-1, Tsukuba 305-8562, Japan
2
SPring8, Japan Synchrotron Radiation Institute (JASRI), Kouto 1-1-1, Sayo-cho, Sayo-gun,
Hyogo 679-5148, Japan
(Received 19 August 2011; accepted 22 September 2011; published online 11 October 2011)
Phase-change GeTe/Sb2Te3 multilayered structures, in which the atomic motion at the layer
interfaces is limited to one dimension, have been shown to require substantially lower switching
energies when compared to monolithic alloys of the same average composition. Here, we report
that in the GeTe/Sb2Te3 superlattice, an extraordinarily large magnetoresistance of DR/R > 2000%
can be induced by application of an electrical field at temperatures exceeding 400 K. This finding
paves the way for development of conceptually new memory devices that combine the merits of
C 2011 American Institute of Physics.
both phase-change and magnetic data storage. V
[doi:10.1063/1.3651275]
Magnetoresistance (MR) is one of the pillars of spinelectronics and is widely applied not only in hard disk
storage but also in nonvolatile solid state memory such as
magnetic random access memory (MRAM). MRAM uses
the tunnel magnetoresistance (TMR) of a magnetic tunnel
junction consisting of ferromagnetic layers interspaced with
an insulating layer such as Al2O3. The use of TMR is essential in down scaling MRAM structures to achieve storage
densities beyond 1 Gbit/in2. (Ref. 1) Although TMR devices
in hard disk applications use Al2O3 as the insulating layer,
recent research has demonstrated that use of a thin MgO
spacer layer can lead to a drastic improvement of magnetoresistance with DR/R > 1000% at 5 K and 500% at room temperature.2,3 Thermal issues related to spin-current injection,
however, present significant challenges to further scaling due
to thermal drift effects resulting in smaller achievable DR/R.
At the same time, phase-change random access memory
(PCRAM) is gaining momentum as the likely candidate technology for the next generation of nonvolatile data storage.
PCRAM devices encode data as structural changes in a thin
film of a (usually) tellurium-based material. In particular, the
stoichiometric compositions along the GeTe/Sb2Te3 pseudobinary tie-line are of great technological interest. Switching
between the disordered and crystalline structural states is
typically achieved through Joule heating.4 Figure 1 (upper
panel) shows schematic representations of the two switching
states in a typical PCRAM device: in the reset state, a highresistance amorphous hemispherical dome covering the bottom electrode forms due to the application of a relatively
high current pulse and subsequent melt-quench; in the conductive set state, the dome reverts to the low-resistance crystalline state by application of a relatively low current and
long duration pulse. The large property changes observed in
the phase-change memory (PCM) between the ordered and
disordered phases have been ascribed to a change in the
nature of the bonding.4–6 For Ge2Sb2Te5, this involves a
a)
Author to whom correspondence should be addressed. Electronic mail:
j-tominaga@aist.go.jp.
0003-6951/2011/99(15)/152105/3/$30.00
change between delocalized resonant bonding in the ordered
phase and localized covalent bonding in the amorphous
phase. While the disordered phase is normally obtained by
quenching the molten phase, in a recent report, multilayered
GeTe/Sb2Te3 structures, referred to as interfacial phasechange memory (iPCM), were found to lead to a similar
change in bonding properties via a solid-solid phase transition.7 In the iPCM structure, the confinement of the switching process to interfaces reduces entropic losses resulting in
an order of magnitude improvement in energy consumption.7
The iPCM structure remains crystalline in both lowresistance (Fig. 1, lower panel) and high-resistance states
and the property contrast is due to a change in the coordination number of Ge atoms.7
Since Ge-Sb-Te alloys do not have a magnetic moment,
until now magnetism was not thought to play a role in
switching dynamics or electrical properties. At the same
time, one of the end points (Sb2Te3) is a topological insulator8 and Ge2Sb2Te5 has also been predicted to be a topological insulator for certain layer sequences.9 It may be
expected, therefore, that application of a magnetic field
might affect material properties, especially for the GeTe/
Sb2Te3 layered iPCM phase as application of an external
magnetic field is expected to further contribute to the Rashba
effect induced splitting in spin states.10 We suspect that the
topological insulating state in the iPCM devices does not
occur only at the surface but is also present at all interfaces
greatly magnifying the magneto-resistance effect and at the
same time, reducing the effects of surface contamination that
would otherwise make fabrication of a device impractical.
iPCM devices with the [(GeTe)2(Sb2Te3)4]8 structure
were grown on 50 nm TiN heater rod arrays prefabricated on
a Si substrate. Prior to film deposition, a 5-nm thick Sb2Te3
film was initially deposited to ensure strong pseudo-cubic
h111i orientation of subsequent layers. Each layer of GeTe
and Sb2Te3 was formed by a helicon-wave sputtering system
with the 2 in. targets. Due to the long target-substrate sputtering distance of 200 mm, atomically controlled layers could
be fabricated by automatically opening and closing shutters
99, 152105-1
C 2011 American Institute of Physics
V
Author complimentary copy. Redistribution subject to AIP license or copyright, see http://apl.aip.org/apl/copyright.jsp
152105-2
Tominaga et al.
Appl. Phys. Lett. 99, 152105 (2011)
FIG. 1. (Color online) Reset (left) and Set (right) states
of PCRAM (top) and iPCM (bottom) devices. The
cylindrical heater acts to heat the PCM film. In
PCRAM, upon melting and subsequent quenching, a
hemispherical amorphous high-resistance dome is
formed above the heater (reset state). Heating the dome
above the crystallization temperature with a relatively
low current and long duration pulse reverts its structure
to that of the low-resistance crystalline phase (set state).
In the iPCM, on the other hand, the transition between
two crystalline states with different bonding characteristics (resonant and covalent in the set and reset states,
accordingly) at interfaces is responsible for the property
contrast.
in front of the targets and of the substrate station. The layer
deposition temperature was fixed at 520 K, taking into
account the crystallization temperature (500 K) of GeTe. The
thicknesses of the layers in the iPCM structure were 0.4 nm
for GeTe and 1.0 nm for Sb2Te3. A cross sectional image of
the device by high-resolution TEM has been published by
Tominaga et al.11
Figure 2 compares the switching behavior of a composite
Ge1Sb4Te7 device and an identical composition iPCM device
with the [(GeTe)2(Sb2Te3)4]8 iPCM structure. As expected for
a non-magnetic material, the electrical switching properties of
a PCRAM device based on a composite Ge1Sb4Te7 alloy were
insensitive to the application of a magnetic field of 0.1 T
(Fig. 2(a)). The current (I)- voltage (V) characteristics the
iPCM device, however, showed a strong dependence on a
magnetic field (Fig. 2(b)) applied in-plane. Prior to the application of a magnetic field, the iPCM device characteristics
were similar to those of the composite Ge1Sb4Te7 device with
switching from the reset to the set state occurring at 0.85 V
and 0.01 mA, (Joule heat released under such conditions corresponds to temperatures ca. 430 K).12 Upon application of a
magnetic field (0.1 T), however, the threshold voltage
increased to 2.0 V (corresponding to temperatures ca.
570 K12) and the slope of the I-V curve changed. The effect of
the magnetic field was completely reversible. It should be
noted that the I-V curve for the iPCM (Fig. 2(b)) at voltages
above Vset has reproducible deviations from the expected
straight-line behavior observed for the composite material that
are reminiscent of the quantum spin Hall effect.
Figure 3 shows resistance (R)—current (I) plots corresponding to the measurements shown in Fig. 2. To characterize the phase-change behavior, R-I plots are typically
measured between switching pulses at a constant voltage of
0.5 V during an I-V curve measurement. The composite
Ge1Sb4Te7-based PCRAM device showed essentially indistinguishable U- or W-shape curves with switching occurring
in a resistance range between 4 MX and 40 KX both with
and without a magnetic field. On the other hand, the resistivity of the iPCM device in the presence of a magnetic field
did not show any change remaining pinned at around 4 MX,
while the two curves for the case without a magnetic field
(before and after) clearly showed a nonvolatile memory
switch between 4 MX and 200 KX. It is clear from this result
that not only the application of a magnetic field results in a
significant increase in the threshold voltage but also modifies
the response of the phase change device from that of a
memory cell to that of a threshold switch (that reverts to its
original stage once the voltage is decreased).13 The magnetoresistance over the voltage range of 1 to 2 V reaches the
value of DR=R & 2000%. Alternatively, the observed effect
FIG. 2. (Color online) Current-Voltage characteristics of a PCRAM device fabricated using composite Ge1Sb4Te7 (left) and an identical composition iPCM
device (right). As the voltage reaches the threshold value of Vset, the device switches into the low-resistance state. Device characteristics prior to application of
a magnetic field, with an external magnetic field of 0.1 T applied, and after removal of the field are shown. The composite device (left) did not show any
change among the three states, while the iPCM device (right) clearly showed a voltage shift under the magnetic field, indicated as Vset(mag.). A sequence of
300 ns pulses was used for the measurements. Two scans for each state are shown to demonstrate reproducibility. Dots are experimental data and solid lines
are guides for the eye.
Author complimentary copy. Redistribution subject to AIP license or copyright, see http://apl.aip.org/apl/copyright.jsp
152105-3
Tominaga et al.
Appl. Phys. Lett. 99, 152105 (2011)
FIG. 3. (Color online) Resistance (R)-Current (I) plots of a PCRAM device (left) and an iPCM device (right) measured at 0.5 eV. Device characteristics prior
to application of a magnetic field, with an external magnetic field of 0.1 T applied, and after removal of the field are shown. The composite device did not
show any change among the three states, while the iPCM device showed a high resistance under the magnetic field (upper line). Two scans for each state are
shown to demonstrate reproducibility
may be interpreted as a magnetic-field induced stabilization
of the high-resistance (amorphous) phase. This result also
demonstrates that the use of a magnetic field can serve to
“write protect” an iPCM device.
It is speculated that the unusually large electrical fieldinduced magnetoresistance in the iPCM device arises from a
spin current resulting from the topological insulating nature of
the multilayered structure where the role of interfaces is
enhanced due to its multilayered nature. In topological insulators,
the bulk is a conventional insulator. However, time-reversal symmetry-protected conducting states exist on the surface (at interfaces) which are resistant to scattering due to the locking of spin to
the propagation direction and destructive quantum mechanical
interference. The coupling of the spin to the momentum in the
absence of a magnetic field originates from quantum mechanical
spin-orbit coupling (SOC), which is larger for heaver elements.
This coupling in conjunction with the topological features of the
band structure, specifically the band inversion at k ¼ 0, leads to
predictions of metallic behavior on the h111i surface even in the
absence of a magnetic field for materials such as Sb2Te3,
Bi2Te3,14 and leads to up and down spin current channels flowing along surface edges. A similar magnetoresistance effect was
observed at ca. 4 K and B ¼ 3.0 T, in the layered organic superconductor (bis(ethylenedithio)tetrathiafulvalene (BEDT-TTF),
(BEDT – TTF)2I3) with iodine intercalated between the layers, a
material that exhibits topological insulator properties.15
All electron calculations using the LO-LAPW code
Wien2K16 demonstrated that the electronic structure of
iPCM is typical of a metal if SOC is neglected but changes
drastically, (exhibiting Dirac cone-like behavior at k ¼ 0 and
inversion symmetry in the presence of SOC. These results
suggest that the iPCM structure can act as a topological insulator due to the presence of a strong Rashba effect. For the
iPCM structures reported here, calculations show a Rashba
effect induced spin-splitting of more than 200 meV, a value
six times greater than that expected from thermal fluctuations
at room temperature. These results strongly suggest that the
topological nature of the structure is preserved at higher
temperatures.
The observed strong magnetic response of iPCM at temperatures exceeding room temperature paves the way for
designing conceptually novel memory materials with the
combined merits of phase-change and magnetic memories.
The observed effect may also serve as the basis for a new
type of magnetic sensor that can operate at high temperatures. Such a device would match well the requirements for
the thermally assisted magnetic heads being proposed now in
the literature.17
This research was funded by the Japan Society for the
Promotion of Science (JSPS) through the Funding Program
for World-Leading Innovative RD on Science and Technology (FIRST Program), initiated by the Council for Science
and Technology Policy (CSTP). We thank Elpida Memory,
Inc. for discussions on the device measurements and Ms. R.
Kondo for her assistance in device fabrication.
1
C. Chappert, A. Fert, and F. N. V. Dau, Nature Mater. 6, 813 (2007).
S. Yuasa, T. Nagahama, A. Fukushima, Y. Suzuki, and K. Ando, Nature
Mater. 3, 868 (2004).
3
Y. M. Lee, J. Hayakawa, S. Ikeda, F. Matsukura, and H. Ohno, Appl.
Phys. Lett. 90, 212507 (2007).
4
M. Wuttig and N. Yamada, Nature Mater. 6, 824 (2007).
5
A. Kolobov, P. Fons, A. Frenkel, A. Ankudinov, J. Tominaga, and T.
Uruga, Nature Mater. 3, 703 (2004).
6
A. V. Kolobov, M. Krbal, P. Fons, J. Tominaga, and T. Uruga, Nature
Chem. 3, 311 (2011).
7
R. E. Simpson, P. Fons, A. V. Kolobov, T. Fukaya, M. Krbal, T. Yagi, and
J. Tominga, Nature Nanotech. 6, 501 (2011).
8
D. Hsieh, Y. Xia, D. Qian, L. Wray, F. Meier, J. H. Dil, J. Osterwalder, L.
Patthey, A. V. Fedorov, H. Lin, A. Bansil, D. Grauer, Y. S. Hor, R. J.
Cava, and M. Z. Hasan, Phys. Rev. Lett. 103, 146401 (2009).
9
J. Kim, J. Kim, and S.-H. Jhi, Phys. Rev. B 82, 201312 (2010).
10
Y. Zhang, K. He, C.-Z. Chang, C.-L. Song, L.-L. Wang, X. Chen, J.-F. Jia,
Z. Fang, X. Dai, W.-Y. Shan, S.-Q. Shen, Q. Niu, X.-L. Qi, S.-C. Zhang,
X.-C. Ma, and Q.-K. Xue, Nature Phys. 6, 584 (2010).
11
J. Tominaga, R. Simpson, P. Fons, and A. Kolobov, in Proceedings of the
European Symposium on Phase Change and Ovonic Science (E\PCOS,
Milano, Italy, 2010), pp. 54–59.
12
J. Tominaga, P. Fons, A. Kolobov, T. Shima, T. C. Chong, R. Zhao, H. K.
Lee, and L. Shi, Jpn. J Appl. Phys. 47, 5763 (2008).
13
R. R. Shanks, J. Non-Cryst. Solids 2, 504 (1970).
14
H. Zhang, C.-X. Liu, X.-L. Qi, X. Dai, Z. Fang, and S.-C. Zhang, Nature
Phys. 5, 438 (2009).
15
N. Tajima, S. Sugawara, R. Kato, Y. Nishio, and K. Kajita, Phys. Rev.
Lett. 102, 176403 (2009).
16
K. Schwarz and P. Blaha, Comput. Mater. Sci. 28, 259 (2003).
17
A. Moser, K. Takano, D. Margulies, M. Albrecht, Y. Sonobe, Y. Ikeda, S.
Sun, and E. Fullerton, J. Phys. D: Appl. Phys. 35, R157 (2002).
2
Author complimentary copy. Redistribution subject to AIP license or copyright, see http://apl.aip.org/apl/copyright.jsp